Optimized Sales Optimized Marketing Target Accounts For CROs For CFOs For CMOs Blog News Glossary Compare Tools About Schedule a Demo
Sales Forecasting

How to Structure a Revenue Forecast Spreadsheet

Pete Furseth 6 min read
forecast modelingsales forecastingrevenue operationsrevenue forecasting
How to Structure a Revenue Forecast Spreadsheet
Home/ Blog/ How to Structure a Revenue Forecast Spreadsheet

Most forecast spreadsheets fail on architecture rather than math. The formulas are fine. The problem is that assumptions are buried inside cells, the calculation logic is scattered across tabs, and nobody can trace the final number back to the inputs that produced it. Structure the file properly and the same math becomes defensible in a board meeting.

What tabs does a forecast model need?

Five, each with one job and a strict rule about what may write to it.
TabContainsRule
RawCRM export of open and closed opportunitiesImport only. No manual edits, ever
AssumptionsEvery judgment call, one per labeled cellNo values hardcoded anywhere else
RatesConversion, cycle length, and realization by segmentCalculated from Raw, never typed
EngineThe forecast calculation, row per segment or periodReferences other tabs only
OutputThe number, the bridge, and the variance viewFormulas point at Engine, no logic of its own
The discipline that matters is directional flow. Data moves left to right through those tabs and never backward. When someone patches an output cell because the total looked wrong, the model stops being a model and becomes a spreadsheet with a preferred answer in it.
Put this to work on your numbers
Run your own numbers with the free Forecast Accuracy Scorecard, then see how ORM builds it into a custom model.

Where should assumptions live?

On one tab, one per cell, each labeled with its source and the date it was last reviewed.

Every forecast contains judgment. What share of pipeline is realistic. How much of the quarter comes from deals that do not exist yet. Whether a large deal is included at full value. Those calls are legitimate. Hiding them inside a formula is not.

Give each assumption three columns: the value, where it came from, and when it was last checked. The source column is what makes the review productive. An assumption sourced to "last four quarters of closed-won" invites a different conversation than one sourced to "CRO estimate," and both deserve to be in the model as long as the label is honest.

How do you build the rate library?

Calculate every rate from the raw tab with formulas, so a data refresh updates the rates automatically.

The rate library holds stage conversion, average cycle length, and the realization ratio between pipeline amount and closed-won amount, each split by segment. These are the numbers that do the actual forecasting work.

Typing a rate is the most common error in a homemade model. The rate was accurate when someone calculated it, then the market moved and nobody recalculated. Formula-driven rates that recompute on refresh avoid the slow decay that makes a model look reliable for two quarters and useless in the third. Pipeline amounts also run above closed-won amounts, sometimes by a factor of two, so the realization ratio belongs in this library rather than in someone's head.

How do you stop the model from breaking every quarter?

Drive all period logic from a single date cell and let columns calculate their own labels.

Hardcoded quarter headers are the leading cause of broken models. Someone rolls the file forward, updates six of eight headers, and the SUMIFS quietly pull the wrong range. The output still produces a number, which is why the error survives.

Put the period start date in one cell on the Assumptions tab. Every column header calculates from it. Every date filter references it. Rolling forward becomes a single edit that propagates through the file, and the reconciliation against last period's file becomes possible because the structure did not change.

Two more habits protect the file. Name the ranges you reference more than twice, so a formula reads in English rather than in coordinates. Keep one row per deal or per segment rather than merging cells for presentation, because merged cells destroy the ability to filter and sort the data underneath.

How do you make the output auditable?

Every number on the output tab should trace to a raw record or a labeled assumption in two clicks.

Traceability is what separates a forecast from an opinion with formatting. When a board member asks why the number moved by $400,000 since last month, the answer needs to arrive in the meeting rather than in a follow-up email three days later.

Build a bridge on the output tab rather than a single total. Show last period's forecast, then the movement from new pipeline created, deals won, deals lost, deals pushed to a later period, and amount changes on existing deals. The bridge answers the question everyone actually asks, which is what changed and why.

The same principle applies to AI-assisted analysis. If a model produces slides without pointing back to the point of truth behind each number, validating those numbers costs as much time as building the deck yourself. Traceability is the requirement, whether the calculation happened in a cell or in a model.

When does a spreadsheet stop being the right tool?

When the refresh cycle runs slower than the decision cycle, or when more than two people need to edit it.

A quarterly planning model can live in a spreadsheet indefinitely. A weekly operating forecast usually cannot, because the manual assembly cost grows with the number of segments, and the file becomes a bottleneck attached to one person.

The tell is timing. If your forecast is accurate but arrives in week eleven of a thirteen week quarter, the accuracy bought you nothing, because the quarter already happened. The value of a forecast is knowing the shape of the period early enough to change it. Our writeup on how to forecast revenue covers the methods that scale past a spreadsheet, and the forecast accuracy glossary entry defines the measures you should be tracking against whichever tool you use.

Before you migrate, fix the structure. A team that moves a badly organized spreadsheet into software gets the same confusion at higher cost. Getting the tab architecture right first also tells you exactly which logic needs to be rebuilt, which makes the migration a translation exercise rather than a redesign. Additional habits worth adopting are collected in sales forecasting best practices.

Frequently Asked Questions

How many tabs should a revenue forecast model have?

Five is enough for most B2B SaaS teams: raw data, assumptions, rate library, calculation engine, and output. Models with twenty tabs are usually hiding calculation logic in places nobody will find during a quarter close, which is exactly when someone needs to find it.

Where should assumptions live in a forecast model?

On a single tab, each in its own labeled cell, with a source note beside it. No assumption should appear inside a formula. If a reviewer has to open a formula to discover you assumed a 22 percent win rate, the model is not auditable and the number will not survive scrutiny.

Should the forecast spreadsheet contain raw CRM data?

Yes, on a dedicated tab that nothing writes to except the import. Calculations reference it, never modify it. The moment someone edits a value on the raw tab to correct a deal, the model stops matching the system of record and every reconciliation afterward becomes an argument.

How do you keep a forecast model from breaking each quarter?

Drive every period reference from one date cell rather than typing quarter labels into headers. Columns calculate their own period from that cell. Rolling the model forward becomes a one-cell change instead of a rebuild, which is where most spreadsheet errors are introduced.

When should you stop using a spreadsheet for forecasting?

When the model needs deal-level history across thousands of opportunities, when more than two people edit it, or when the refresh cycle is slower than the decision cycle. A spreadsheet that takes three days to update cannot inform a weekly forecast call.

PF
Pete Furseth
ORM Technologies
Pete has built custom revenue forecast models for B2B SaaS companies for over a decade.

See how ORM turns these insights into action

ORM builds custom revenue forecast models for B2B SaaS companies. Not dashboards. Prescriptive analytics that tell you what to do next.

Schedule a Demo